On Tue, Dec 23, 2014 at 8:48 AM, L. David Baron <dba...@dbaron.org> wrote:
> On Tuesday 2014-12-23 08:36 -0800, Eric Rescorla wrote: > > > Why not pass the raw pointer to the function? > > > > > > My general theory is that smart pointers, once boxed, should never be > > unboxed. > > The major arguments I see for unboxing is performance, but if you pass a > > ref, > > then you don't have the increment/decrement cycle. > > > > > > > > > That's what we do in most of the places in the code. > > > > > > > Yes, I think this is unwise. > > Our convention has always been to pass raw pointers, generally with > the assumption that the caller is expected to ensure the pointer > lives across the function call. > > Passing references to smart pointers doesn't work if you need > implicit conversions from derived class to base class (e.g., if the > function takes Base*, and the caller has nsRefPtr<Derived>), which > is pretty important in many areas of our codebase. > Hmm... This seems to work fine with RefPtr, but it *doesn't* work with nsRefPtr (see end of message for the test code). I'm guessing because of: http://dxr.mozilla.org/mozilla-central/source/mfbt/RefPtr.h?from=RefPtr&case=true#239 Passing smart pointers (non-references) would cause tons of extra > reference count traffic, which is particularly problematic when the > reference-counting functions are virtual and/or threadsafe. Right, that's the reason for using references. -Ekr class Base { public: NS_INLINE_DECL_THREADSAFE_REFCOUNTING(Base); private: ~Base() {} }; class Derived : public Base { }; void base_ref(const RefPtr<Base>& b) { } void base_ptr(Base* b) { } void foo() { RefPtr<Base> base; RefPtr<Derived> derived; base_ref(base); base_ref(derived); base_ptr(base); base_ptr(derived) } _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
